*** elo has quit IRC | 00:01 | |
*** elo has joined #akanda | 00:02 | |
*** elo has quit IRC | 00:07 | |
*** swinn has joined #akanda | 00:25 | |
adam_g | keystone v3 / auth sessions lookin good now https://review.openstack.org/#/c/232247/ | 00:44 |
---|---|---|
*** cleverdevil has quit IRC | 00:46 | |
*** swinn has quit IRC | 01:19 | |
markmcclain | adam_g: sadly master again | 02:48 |
openstackgerrit | Merged stackforge/akanda-rug: Add driver framework https://review.openstack.org/215227 | 03:21 |
adam_g | oh man. so many patches need rebasing now | 03:54 |
adam_g | markmcclain, any luck /w the port thing? im starting a fresh env now | 03:55 |
markmcclain | yeah... I think the cause is me being brain dead from waking up too early | 03:57 |
adam_g | markmcclain, as in the stuff isnt really working or PEBAC? | 03:58 |
markmcclain | I'm waiting for a rebuild to finish | 03:58 |
adam_g | PEBCAK rather | 03:59 |
markmcclain | PEBCAK | 03:59 |
adam_g | ah ok. ya, that fix worked great yesterday when i applied it | 04:01 |
adam_g | re: kilo, does the API support disabling port security via the port API or do we need to do something different there? | 04:01 |
markmcclain | so the test case I thought should not fail should fail when things are working right | 04:02 |
markmcclain | I'm going to +A | 04:06 |
* adam_g sips beer slowly and parses that | 04:06 | |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Track last_boot from nova server not instance_info creation https://review.openstack.org/213951 | 04:39 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Add support for hash-based RUG scale out https://review.openstack.org/195366 | 04:39 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Convert to using Keystone sessions for Keystone V3 support https://review.openstack.org/232247 | 04:40 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Convert to using Keystone sessions for Keystone V3 support https://review.openstack.org/232247 | 04:48 |
*** stanchan has joined #akanda | 05:05 | |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Properly detect backing instance is gone https://review.openstack.org/220345 | 05:10 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Delete VRRP and MGT ports on router delete https://review.openstack.org/219951 | 05:10 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Enrich functional test suite https://review.openstack.org/219952 | 05:11 |
*** ronis has joined #akanda | 07:10 | |
*** ronis has quit IRC | 08:53 | |
*** ronis has joined #akanda | 08:57 | |
*** ronis has quit IRC | 09:42 | |
*** ronis has joined #akanda | 10:52 | |
*** ronis_ has joined #akanda | 11:38 | |
*** ronis__ has joined #akanda | 11:40 | |
*** ronis has quit IRC | 11:41 | |
*** ronis_ has quit IRC | 11:43 | |
*** ronis__ has quit IRC | 12:30 | |
*** ronis has joined #akanda | 13:08 | |
*** puranamr has joined #akanda | 13:22 | |
*** puranamr has quit IRC | 13:41 | |
*** ronis has quit IRC | 15:23 | |
*** ronis has joined #akanda | 16:46 | |
*** cleverdevil has joined #akanda | 16:54 | |
*** cleverdevil has quit IRC | 16:54 | |
*** cleverdevil has joined #akanda | 16:55 | |
*** cleverdevil has quit IRC | 17:06 | |
*** cleverdevil has joined #akanda | 17:07 | |
*** cleverdevil has quit IRC | 17:08 | |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Convert to using Keystone sessions for Keystone V3 support https://review.openstack.org/232247 | 17:14 |
*** cleverdevil has joined #akanda | 17:16 | |
*** cleverdevil has quit IRC | 17:17 | |
ryanpetrello | so Mark and I figured out the tenant network thing | 17:29 |
ryanpetrello | it's a security group thing | 17:29 |
ryanpetrello | if you add a rule to your cirros (tenant) VM's security group | 17:29 |
ryanpetrello | e.g., | 17:29 |
ryanpetrello | https://ryanp.dev.newdream.net/co.py/10-09-2015-13-29-35-39c08c24-84f7-4828-9d2f-299065d4d5ac.png | 17:29 |
ryanpetrello | bah, working link: | 17:30 |
ryanpetrello | https://screens.objects.dreamhost.com/10-09-2015-13-29-35-39c08c24-84f7-4828-9d2f-299065d4d5ac.png?Signature=JlI%2Bomn4upMlY3eRUTiAV6VdQOo%3D&Expires=1444498199&AWSAccessKeyId=8Y1FmVQdlqLchxaCahBd | 17:30 |
ryanpetrello | if you do something like this, TCP starts working from router -> tenant VM | 17:30 |
ryanpetrello | same with ICMP | 17:30 |
ryanpetrello | the upstream security group default that's set up for the demo tenant has v4 and v6 ingress wide open, but restricted to the customer tenant's security group as the remote | 17:31 |
ryanpetrello | so this doesn't work since the router VM is on another tenant/security group | 17:31 |
ryanpetrello | if you change the rule to use a cidr for the private network, it works | 17:31 |
ryanpetrello | this def. seems unique to akanda given the "VM in another tenant" bit | 17:31 |
ryanpetrello | my thinking is that we should update the akanda devstack plugin to actually create a rule like this (restricted to the private cidr) in the demo tenant's security group | 17:32 |
ryanpetrello | adam_g markmcclain thoughts ^ ? | 17:32 |
ryanpetrello | perhaps in https://github.com/stackforge/akanda-rug/blob/master/devstack/plugin.sh#L303 ? | 17:33 |
*** elo has joined #akanda | 17:33 | |
ryanpetrello | instead of opening this special DHCP hole, it might make more sense to just totally open up v4 and v6 across the tenant cidr (for devstack) | 17:35 |
*** cleverdevil has joined #akanda | 17:49 | |
*** cleverdevil has quit IRC | 17:51 | |
*** cleverdevil has joined #akanda | 17:54 | |
adam_g | ryanpetrello, yeah, having devstack do that seems reasonable enough. im not sure we need to do anything for the non-developer case tho | 17:56 |
ryanpetrello | sure | 17:56 |
ryanpetrello | I think this really only makes sense from a development perspective | 17:56 |
adam_g | yea | 17:56 |
adam_g | or admin/testing | 17:56 |
ryanpetrello | and for deployers who care | 17:56 |
ryanpetrello | like us, we'd probably enable this by default for tenants in our cluster | 17:57 |
ryanpetrello | restricting to the private cidr | 17:57 |
ryanpetrello | a common thing we do when troubleshooting stuff for folks | 17:57 |
ryanpetrello | is starting at the router | 17:57 |
ryanpetrello | e.g., "can I ping some tenant VM from the router" | 17:57 |
ryanpetrello | so that's sort of how we discovered this | 17:57 |
adam_g | yup | 17:58 |
adam_g | markmcclain mentioned maybe trying to get somethign into neutron upstream to allow this | 17:58 |
adam_g | b/c i've found myself pinging tenant VMs from router network namespaces as a test in non-akanda worlds, too | 17:59 |
adam_g | "everything is a freaking security group problem" | 18:00 |
davidlenwell | lol | 18:00 |
adam_g | https://etherpad.openstack.org/p/akanda-mitaka-planning | 18:01 |
adam_g | patches that need mergin' listed there | 18:01 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-neutron: Add the akloadbalancerstatus API extension https://review.openstack.org/225370 | 18:06 |
ryanpetrello | adam_g, I tinkered with this a bit | 18:08 |
ryanpetrello | with the plugin | 18:08 |
ryanpetrello | https://github.com/ryanpetrello/akanda-rug/commit/0c18463610d9b48a049ff8b373619e10cf997cb8 seems to do it | 18:08 |
ryanpetrello | how do you feel about this as a default for devstack? | 18:08 |
*** cleverdevil has quit IRC | 18:10 | |
adam_g | ryanpetrello, i feel good about that | 18:11 |
markmcclain | yeah.. that would work | 18:11 |
ryanpetrello | k, I'll open a review | 18:11 |
*** cleverdevil has joined #akanda | 18:11 | |
adam_g | http://www.losangelesduilawyer.org/wp-content/uploads/2014/08/james_brown.jpg | 18:11 |
ryanpetrello | dah nah nah nah nah nah nah | 18:11 |
adam_g | ryanpetrello, that actually gets rid of the hard-coded 192.168.0.0 CIDR that ive been meaning to update | 18:11 |
ryanpetrello | \o/ | 18:11 |
markmcclain | there is one side-effect from this change | 18:12 |
markmcclain | if you have two vms on teh same network in different sec groups they will always be able to communicate | 18:12 |
openstackgerrit | Merged stackforge/akanda-rug: Adds an LBAAS driver https://review.openstack.org/225369 | 18:13 |
markmcclain | I'd have to double check, but this might cause a ripple effect and break tempest test | 18:13 |
adam_g | so ryanpetrello | 18:24 |
adam_g | is the fact that we just changed the instance naming in liberty from ak-$uuid to ak-router-$uuid going to blow up tooling you guys have built up around this? | 18:25 |
openstackgerrit | Merged stackforge/akanda-appliance: Remove unused variables https://review.openstack.org/230571 | 18:25 |
ryanpetrello | adam_g: yes :) | 18:29 |
adam_g | maybe we can make it template-able via config | 18:30 |
openstackgerrit | Ryan Petrello proposed stackforge/akanda-rug: By default, add a secgroup rule to permit all private network traffic. https://review.openstack.org/233213 | 18:32 |
adam_g | markmcclain, just confirmed all the lb stuff works fine with the noop logging n-lbaas driver enabled | 18:32 |
ryanpetrello | that would be really helpful :) | 18:32 |
adam_g | markmcclain, only wart there is that the neutron resources show up /w provider=loggingnoop | 18:33 |
ryanpetrello | markmcclain adam_g https://review.openstack.org/233213 | 18:33 |
adam_g | ryanpetrello, +2 | 18:35 |
*** cleverdevil has quit IRC | 19:03 | |
markmcclain | ryanpetrello: was thinking about AKANDA_DEFAULT_SUBNET_CIDR | 19:19 |
clett | adam_g: added a pep0257 plugin to flake8, it'll help :) | 19:20 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-neutron: Add the akloadbalancerstatus API extension and lbaasv2 plugin https://review.openstack.org/225370 | 19:21 |
adam_g | proper extension loading unlocked ^ | 19:21 |
adam_g | clett, cool. apologies if those dont get merged immediately, too much stuff in-flight atm and those will likely cause added conflicts | 19:21 |
* adam_g lunch | 19:21 | |
clett | adam_g: no problem, just learning the collaborative environment here. once i can get in the simplest change, then i can do code contributions. | 19:22 |
ryanpetrello | markmcclain: ...and :)? | 19:30 |
markmcclain | ryanpetrello: you couldnt figure it out? :) | 19:31 |
ryanpetrello | ...huh? | 19:32 |
markmcclain | ryanpetrello: so there's already a cidr value mainline devstack uses: http://git.openstack.org/cgit/openstack-dev/devstack/tree/lib/neutron-legacy#n553 | 19:33 |
ryanpetrello | oh, I didn't know that :) | 19:34 |
ryanpetrello | $FIXED_RANGE ? | 19:34 |
ryanpetrello | (yea, looks that way) | 19:35 |
ryanpetrello | okay, I'll do ps2 | 19:35 |
markmcclain | yeah... I think we might be using fixed_range the wrong way | 19:37 |
ryanpetrello | in our plugin? | 19:39 |
markmcclain | actually it's this hardcoded: https://review.openstack.org/#/c/232193/ | 19:40 |
markmcclain | oops.. wrong like | 19:40 |
markmcclain | https://git.openstack.org/cgit/stackforge/akanda-rug/tree/devstack/plugin.sh#n186 | 19:40 |
markmcclain | ryanpetrello: we could also start using the floating range: http://git.openstack.org/cgit/openstack-dev/devstack/tree/lib/neutron-legacy#n1276 | 19:42 |
*** cleverdevil has joined #akanda | 19:45 | |
markmcclain | ryanpetrello: sorry to make more work | 19:47 |
ryanpetrello | in another commit, maybe :)? | 19:47 |
markmcclain | sure... I'll propose that in a follow up | 19:47 |
ryanpetrello | k | 20:06 |
ryanpetrello | I'm testing w/ FIXED_RANGE | 20:06 |
ryanpetrello | and will update the review | 20:06 |
openstackgerrit | Merged stackforge/akanda-rug: disable port_security for VRRP ports https://review.openstack.org/232193 | 20:22 |
*** ronis has quit IRC | 20:24 | |
openstackgerrit | Ryan Petrello proposed stackforge/akanda-rug: By default, add a secgroup rule to permit all private network traffic. https://review.openstack.org/233213 | 20:35 |
ryanpetrello | markmcclain adam_g ^^ | 20:35 |
ryanpetrello | tested and it seems to work | 20:36 |
adam_g | ryanpetrello, cool | 20:58 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Properly detect backing instance is gone https://review.openstack.org/220345 | 21:05 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Delete VRRP and MGT ports on router delete https://review.openstack.org/219951 | 21:05 |
adam_g | markmcclain, rebases^ | 21:05 |
*** cleverdevil has quit IRC | 21:20 | |
*** cleverdevil has joined #akanda | 21:26 | |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Track last_boot from nova server not instance_info creation https://review.openstack.org/213951 | 22:08 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Add support for hash-based RUG scale out https://review.openstack.org/195366 | 22:08 |
adam_g | davidlenwell, ^ | 22:08 |
openstackgerrit | Adam Gandelman proposed stackforge/akanda-rug: Adds missing resource CLI file https://review.openstack.org/233293 | 22:20 |
adam_g | markmcclain, davidlenwell ^ looks like i forgot to 'git add' in one of the previous patches | 22:21 |
*** elo has quit IRC | 22:38 | |
*** elo has joined #akanda | 22:38 | |
*** elo is now known as help | 23:11 | |
*** help is now known as Guest78155 | 23:11 | |
*** Guest78155 has quit IRC | 23:22 | |
*** elo has joined #akanda | 23:23 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!